| 90 | 1 | 57 |
| 下载次数 | 被引频次 | 阅读次数 |
目的 在药学工作站中增加深度学习功能,利用该功能评价风险因素及改进药学服务质量。方法 筛选2016年6月至2021年6月就诊的宫颈癌患者1480例,收集其临床资料。在原有药学工作站基础上增加3项深度学习功能,即调研、聚类分析和关联规则分析。收集患者临床资料、用药相关资料及不良事件信息,通过聚类分析(基于K-means算法)计算患者无进展生存期(PFS)类别,关联规则分析(基于Apriori算法)计算最低类PFS与患者资料之间的有效强关联规则,得出导致最低类PFS的风险因素,并采用Logistic法加以验证。之后根据风险因素提出有针对性的措施,以提高药学服务质量。结果 通过聚类分析将患者的PFS分为4类,聚类中心分别为14.28、10.37、7.82和4.60。通过关联规则分析,认为长期用药目标和个体化且合理的药物选择的不明确,特别是针对查尔森合并症指数≥4和鳞癌患者用药的培训不足和配伍错误是导致最低类PFS的原因。结论 在药学工作站中增加深度学习,不仅可用于宫颈癌风险因素的筛选,还有助于药学服务质量的提升。
Abstract：AIM To add deep learning function in pharmaceutical workstations, and use this function to evaluate risk factors and improve the quality of pharmaceutical care. METHODS A total of 1 480 patients with cervical cancer from June 2016 to June 2021 were selected and their clinical data were collected. Three deep learning functions were added to the traditional pharmacy workstations: investigation, cluster analysis and association rule analysis. The patient clinical data, medication related data and adverse event information were collected. The categories of progression-free survival(PFS) of patients were calculated through cluster analysis based on K-means algorithm. The effective strong association rules between the PFS of the lowest category and the characteristics of pharmaceutical service were calculated through association rule analysis based on Apriori algorithm. The risk factors leading to the lowest PFS were analyzed to verify the evaluation results. Targeted measures were taken to improve the quality of pharmaceutical care according to the evaluation results. RESULTS The PFS of patients was divided into 4 categories by cluster analysis of 14.28, 10.37, 7.82 and 4.60. Through association rule analysis, unclear long-term medication goals and individualized and reasonable drug selection, especially insufficient training and compatibility errors for patients with Charlson Index ≥ 4 and squamous cell carcinoma, were considered as the reasons for the lowest type of PFS. Logistic regression analysis proved the scientific and reliability of the test results to some extent. CONCLUSION Adding deep learning in pharmacy workstations can not only be used to screen risk factors for cervical cancer, but also help to improve the quality of pharmaceutical care.
[1]ABRAMOWITZ P W. Future directions in specialty pharmacy[J].Am J Health Syst Pharm, 2021, 78(19):1743.
[2]SHANE R, ROUGH S, CHEN D. It’s time for a “beyond-use date”for designating health-system pharmacy as a center of excellence[J]. Am J Health Syst Pharm, 2021, 78(6):527.
[3]CAMPBELL U, CHEN D, RIM M H. Charting a course for healthsystem specialty pharmacy[J]. Am J Health Syst Pharm, 2021,78(19):1792.
[4]郝晓慧,张志林,卢秀荣,等.信迪利单抗静注联合全身化疗治疗复发转移宫颈癌的临床疗效观察[J].山东医药, 2022, 62(5):78.
[5]LIN C H, WANG L, SHI L. AAPred-CNN:accurate predictor based on deep convolution neural network for identification of antiangiogenic peptides[J]. Methods, 2022, 204:442.
[6]CAMMARASANA S, NICOLARDI P, PATANèG. Realtimedenoising of ultrasound images based on deep learning[J].Med Biol Eng Comput, 2022, 60(8):2229.
[7]马静,赵静,马丽丽,等.中老年妇科恶性肿瘤化疗患者合并症负担及症状群的调查分析[J].新疆医科大学学报, 2022, 45(3):345.
[8]BAASLAND I, VIE G?, ROMUNDSTAD P R, et al. Cervical cancer mortality in Norway according to screening attendance and age[J]. Acta Obstet Gynecol Scand, 2022, 101(9):952.
[9]TONG Y, ORANG'O E, NAKALEMBE M, et al. The East Africa Consortium for human papillomavirus and cervical cancer in women living with HIV/AIDS[J]. Ann Med, 2022, 54(1):1202.
[10]TANG X Y, ZHENG P J, LI X Y, et al. Deep6mAPred:a CNN and Bi-LSTM-based deep learning method for predicting DNA N6-methyladenosine sites across plant species[J]. Methods, 2022,204:142.
[11]BELCIUG S. Learning deep neural networks'architectures using differential evolution. Case study:medical imaging processing[J].Comput Biol Med, 2022, 146:105623.
[12]郭慧敏,姜文澜,杜军.基于K-means聚类的医院住院患者特征分析[J].中国病案, 2018, 19(7):57.
[13]陈默,蔡苗,黄阿红,等.基于K-means聚类与支持向量机的大病患者住院费用影响因素与控制策略研究[J].中国医院管理,2019, 39(5):45.
[14]李曙光,赵敏,张秀英,等.关联规则法用于护士给药错误的数据挖掘分析[J].护理学杂志, 2021, 36(13):104.
[15]贺鲜娇,陈英.基于Apriori算法的某肿瘤专科医院315例护理不良事件关联因素分析[J].全科护理, 2021, 19(16):2180.
[16]冯祥,宋统球,华召来,等.基于SERVQUAL量表对上消化道癌筛查患者的服务质量评估[J].中华肿瘤防治杂志, 2022,29(14):1039.
[17]罗家洪,毛勇,黄兴黎,等.医务人员对新农合定点医疗机构服务质量的评价[J].中国卫生质量管理, 2005, 12(5):13.
基本信息:
DOI：10.19577/j.1007-4406.2023.07.003
中图分类号:R737.33
引用信息:
[1]叶伊佳,王晨宇,田龙.深度学习在宫颈癌风险因素评价及药学服务质量改进中的应用[J].中国临床药学杂志,2023,32(07):498-503.DOI:10.19577/j.1007-4406.2023.07.003.
基金信息:
河北省高等学校科学技术研究项目(编号QN2021081)
2023-07-25
2023-07-25